Midsummer Deficit Irrigation of Alfalfa for Water Conservation in the San Joaquin Valley of California

Bali, Khaled M.; Putnam, Daniel; Wang, Dong; Begna, Sultan; Holder, Brady; Mohamed, Abdelmoneim Zakaria; Paloutzian, Luke; Dahlke, Helen E.; Eltarabily, Mohamed Galal

doi:10.1061/JIDEDH.IRENG-10213

Open access

Technical Papers

Sep 24, 2024

Midsummer Deficit Irrigation of Alfalfa for Water Conservation in the San Joaquin Valley of California

Publication: Journal of Irrigation and Drainage Engineering

Volume 150, Issue 6

https://doi.org/10.1061/JIDEDH.IRENG-10213

PDF

Abstract

A four-year research experiment was conducted on sandy loam soil at the University of California Kearney Agricultural Research and Extension Center in Parlier, California, to investigate the effect of midsummer deficit irrigation on alfalfa yield, irrigation water productivity (IWP), and crop water productivity (CWP). The experiment was a randomized block design with two treatments: full and deficit irrigations with three replications. Applied irrigation water was measured using flow meters and soil matric potentials were monitored using watermark soil moisture sensors. Actual evapotranspiration (

E T_{a}

) values were estimated from Tule Technologies stations. The deficit irrigation treatments resulted in 454, 706, 625, and 815 mm of irrigation water savings as compared to the full irrigation treatments in 2019, 2020, 2021, and 2022, respectively. These values represent 30.3%, 40.9%, 37.0%, and 49.1% of the applied water savings. Alfalfa yield in the deficit treatments was reduced by 3.94, 2.04, 1.25, and

0.40 Mg {ha}^{- 1}

; the equivalent of 18.1%, 11.1%, 7.1%, and 3.0% of the yield for the full irrigation treatment for the four years: with an average reduction of 10.7%. IWP was higher when deficit irrigation was implemented and resulted in 17.09, 16.11, 15.40, and

15.54 kg {ha}^{- 1} {mm}^{- 1}

, in 2019, 2020, 2021, and 2022, respectively. The production function using applied irrigation water (IW, mm) was:

Y (yield in Mg {ha}^{- 1}) = 0.50 \times {(IW)}^{2} - 1,633.75 \times (IW) + 1,338,472

and

Y = - 0.1137 \times {(IW)}^{2} + 233.55 \times (IW) - 103,036

for the full and deficit irrigation treatments, respectively. CWP was 18.6, 16.4, 14.9, and

12.3 kg {ha}^{- 1} {mm}^{- 1}

for fully irrigated treatments, and 15.2, 14.9, 14.3, and

12.6 kg {ha}^{- 1} {mm}^{- 1}

for the deficit irrigation treatments, for 2019, 2020, 2021, and 2022, respectively. Results from this work provide growers with viable deficit irrigation practices that could be implemented during drought periods.

Introduction

Alfalfa is one of the major field crops in California with relatively high water use due to its long growing season (UC Davis 2023). Alfalfa is an important forage legume crop in many parts of the world for dairy and cattle feeding industries. Alfalfa is an integral part of the dairy industry in the state and contributed $1.1 billion to California’s economy in 2021, making it the state’s 13th most valuable crop. Alfalfa has been one of the valuable inputs to California’s dairy industry, which is California’s leading agricultural commodity in cash receipts valued at $7.6 billion in 2021 (CDFA 2022). Alfalfa seasonal applied water ranges from 490,000 to 677,000 ha-m per year (Hanson et al. 2007) due to its relatively long growing season in California. Alfalfa in general has a relatively high water use efficiency (WUE) and deep root system. Irrigating alfalfa in California is important for maintaining a high marketable yield and quality. Practical irrigation and water conservation strategies are needed to increase the WUE of alfalfa, especially during dry years (Alam et al. 2002; Hanson et al. 2007). Underirrigated alfalfa can often cause a significant yield loss (Shewmaker et al. 2011; Anower et al. 2015), while overirrigation causes loss of alfalfa stands (Bai and Li 2003; Putnam et al. 2017a), and it can lead to root waterlogging that decreases growth and yield (Barta and Sulc 2002).

The impact of the recent drought in California and declining groundwater levels could be partially mitigated by implementing water conservation practices such as deficit irrigation on alfalfa. Deficit irrigation is a water application practice that applies less water than full-crop evapotranspiration (Fereres and Soriano 2007). Intentional reduction in crop water use through deficit irrigation practices is an efficient practice for conserving water resources if it is economically managed (Perez-Blanco et al. 2020). Alfalfa is a drought-tolerant crop and a good candidate for the implementation of deficit irrigation practices. Midsummer deficit irrigation could be an option to conserve water during drought years or to transfer water to other water shortage regions (Hanson et al. 2007; Lindenmayer et al. 2011; Ottman and Putnam 2017).

Deficit irrigation affects alfalfa yield and possibly alfalfa feed quality, and it is important to have a balance between them. Deficit irrigation may improve alfalfa hay quality and save a significant amount of water that could offset the reductions in yield (Lindenmayer et al. 2011; Ismail and Almarshadi 2013). The primary factors that typically affect feed quality are plant maturity at harvest, weed management, temperature, and harvesting methodology. Secondary factors include variety, soil type, fertility, irrigation management, and insect and disease damage. In addition, biotechnological traits may have impacts on quality as well. Harvest scheduling is the single most important practice impacting alfalfa hay quality. Quality should never be considered in isolation from yield goals due to the tradeoff between yield and quality that is observed in forage crops (Putnam and Orloff 2016).

Alfalfa yield and crop evapotranspiration relationships vary by growing region due to different climates (Hanson et al. 2007). Different relationships between yield and applied irrigation amount were previously documented by various researchers. A linear relationship was found between alfalfa yield and the amount of applied irrigation in several studies (Shewmaker et al. 2011; Klocke et al. 2013; Rogers et al. 2016; Li and Su 2017). Djaman et al. (2020) found a curvilinear relationship between yield and applied irrigation amount, while a third-order polynomial relationship was reported by Yang et al. (2019). Slama et al. (2011) highlighted the importance of utilizing deficit irrigation in alfalfa for water conservation. Over a two-year experiment conducted in the San Joaquin Valley, California, alfalfa yield decreased by 65% to 71% in the early-summer deficit treatment where there was no irrigation after June until the following spring, compared with the fully irrigated treatment (Frate et al. 1991). Alfalfa yield was reduced to 46% under midsummer deficit irrigation as compared to those of fully irrigated treatments in the Palo Verde Valley of southern California (Putnam et al. 2000).

Alfalfa is planted on approximately 235,000 ha in California, and the eight SJV counties had more than 99,000 ha of alfalfa valued at $445 million in 2021 (CDFA 2022). The total water use on alfalfa in the SJV in 2021 was

1.38 billion m^{3}

(DWR 2023). The SJV’s 1.82 million ha of cropland needed about

19.7 billion m^{3}

of applied water (AW) in 2018. Tree and vine crops took 69% of the AW, 11% of the AW went to alfalfa and pastures, 7% to corn and other silage crops, 8% to other field and grain crops, and 5% went to vegetables and nontree fruits (Escriva-Bou et al. 2023).

Few studies have been done on midsummer deficit irrigation in the San Joaquin Valley. The objective of this study was to evaluate and document the impact of the midsummer deficit irrigation on alfalfa yield and productivity. Profitable and sustainable alfalfa forage production systems with high IWP are needed for conserving water resources in the San Joaquin Valley of California and in many water-short semiarid and arid regions of the western US. Results of this study will provide growers in the region with tools to employ midsummer deficit irrigation practices to address challenges related to drought and limited water supplies in California and other similar production regions.

Material and Method

Study Area and Experiment Design

The study was conducted at a research field located at the University of California, Kearney Agricultural Research and Extension Center (KARE), Parlier, California (36°36′13.1′′N, 119°30′39.1′′W), between 2019 and 2022 [Fig. 1(a)]. The soil at the location is Hanford sandy loam (coarse-loamy, mixed, superactive, nonacid, thermic Typic Xerorthents), a well-drained soil, (Web Soil Survey, Web Soil Survey, n.d.) (Table 1). The alfalfa cultivar AmeriStand 835NTS RR (Fall Dormancy 8) was planted in October 2018 at a rate of

34 kg {ha}^{- 1}

. Preplant monoammonium phosphate fertilizer (11-52-0) was applied at a rate of

168 kg {ha}^{- 1}

. A potassium sulfate fertilizer was also applied at a rate of

224 kg {ha}^{- 1}

. These dosages are consistent with the common alfalfa fertilization used in the region (CDFA 2024). Sprinkler irrigation was used for germination, and then surface irrigation was used during the entire experiment. The 1.53 ha field was divided into 12 border strips with a 1.5 m border at 250 mm height between strips, each strip was 16.5 m wide and 85 m long except strips 11 and 12 that were only 8.25 m wide (just to keep the same number of replicates for each irrigation treatment) [Fig. 1(b)].

Fig. 1. (a) Experimental field at UC KARE Center; and (b) experimental layout with a randomized complete block design (image data © 2023 Google).

Table 1. Physical and chemical properties of the soil at the experimental site (Web Soil Survey, n.d.)

Soil layer (m)	Soil texture			Organic matter (%)	pH	Available water ( ${cm}^{3} {cm}^{- 3}$ )	Bulk density ( $g {cm}^{- 3}$ )
Soil layer (m)	Clay (%)	Sand (%)	Silt (%)	Organic matter (%)	pH	Available water ( ${cm}^{3} {cm}^{- 3}$ )	Bulk density ( $g {cm}^{- 3}$ )
0–0.3	12.5	68	19.5	0.75	7.6	0.13	1.47
0.3–1.5	12.5	68	19.5	0.33	7.7	0.13	1.45

The experiment was randomized block design with three replications where two irrigation treatments were implemented: full irrigation treatment (two surface irrigation events per cutting over the growing season), and deficit irrigation (two surface irrigation events per cutting until the 5th cut, early August, in the season; then, no IW was applied after the early August cutting). This is one of the two common practices implemented on alfalfa by California growers to achieve water savings. The other practice is complete land fallowing for a year or more. Both of these practices are used to generate conserved water for transfer to urban areas or to other crops that cannot survive with deficit irrigation. Since surface irrigation is the common method of irrigation on alfalfa, it is not practical to deficit irrigate during the growing season. Deficit irrigation during the season could be implemented if pressurized irrigation systems are used. Irrigation scheduling was conducted using the standard commercial growing practices for alfalfa production on relatively light soils. Two irrigations per cutting were scheduled based on ET_a estimates from Tule Technologies, soil moisture conditions in the field, and the required cutting intervals to mimic standard practices in the region. Water was delivered to the border strips from gated aluminum pipes, and the water volume was determined using McCrometer propeller flow meters in a 6-inch pipe (McCrometer 2023). Flowmeters were last calibrated in January 2021 and had an accuracy level of 1.5%. Water applied to each check was metered during the growing season.

Climate Data

Daily meteorological data were collected from the California Irrigation Management Information System (CIMIS, weather station No. 39, Parlier/Fresno County) located at the research center (State of California, n.d.). These data for solar radiation,

maximum / minimum / average

air temperature, average relative humidity, wind speed, and soil temperature were used to calculate the reference evapotranspiration via the Penman–Monteith equation. The actual evapotranspiration was recorded by the surface renewal method (Tule Sensors 2024), and the crop coefficient (Kc) was calculated and subsequently compared with the values in Table 12 of FAO 56.

K c

equals 0.40, 1.20, and 1.15 for alfalfa representing immediately following cutting, at full cover, and immediately before cutting, respectively, where the growing season is described as a series of individual cutting cycles. The area has a Mediterranean climate with an average annual precipitation of about

200 mm {year}^{- 1}

. The total annual precipitation, all in the form of rainfall, was 268, 149, 231, and 131 mm in 2019, 2020, 2021, and 2022, respectively. Similar trends of daily air temperature, solar radiation, and average daily wind speed were observed over the four years of study (Fig. 2). The daily average temperature ranged from 3.3 to 33°C over the study period. Wind speed was lower than

5.5 m s^{- 1}

. A monthly average metrological data for the study period is shown in Table 2.

Fig. 2. (a) Maximum, minimum, mean air temperature, and precipitation on a weekly basis; and (b) weekly average solar radiation and wind speed.

Table 2. Monthly mean climate data for the experimental site, Parlier, CA, US

Year	Month	Total, $E T_{o}$ (mm)	Precipitation (mm)	Average air temperature (°C)	Average wind speed ( $m / s$ )	Average dew point (°C)	Average solar radiation ( $W m^{- 2}$ )
2019	January	37.6	37.7	9.8	1.4	7.7	105
	February	48.7	79.7	8.8	2.0	5.7	143
	March	97.2	26.1	12.9	1.7	7.9	216
	April	147.6	4.0	18.3	1.8	10.2	276
	May	154.9	54.2	18.2	2.0	10.6	283
	June	211.8	0.0	25.6	2.0	13.2	332
	July	220.8	0.0	27.2	1.9	13.9	338
	August	202.4	0.0	27.2	1.7	15.3	313
	September	149.5	0.0	22.8	1.7	12	260
	October	102.7	0.0	15.5	1.3	6.4	207
	November	60.1	13.4	11.5	1.3	5.2	135
	December	28.8	53.0	9.4	1.3	7.3	83
2020	January	36.7	16.3	8.2	1.3	5.7	100
	February	73.5	0.2	10.8	1.4	4.0	187
	March	87.6	69.0	12.5	1.7	7.0	194
	April	130.5	29.8	16.9	1.9	10.1	261
	May	197.8	8.3	21.5	2.1	9.9	335
	June	212.1	0.0	24.9	2.1	11.9	345
	July	222.2	0.0	27.1	1.8	13.4	348
	August	189.5	0.0	27.6	1.6	15.8	287
	September	121.6	0.2	23.2	1.2	14.5	211
	October	93.8	0.0	18.5	1.1	10.8	177
	November	53.7	6.9	10.3	1.2	5.4	135
	December	34.0	18.1	7.5	1.3	1.0	100
2021	January	46.2	62.0	8.8	1.5	4.4	107
	February	63.9	5.2	10.6	1.5	5.2	174
	March	103.3	22.7	12	1.7	4.9	230
	April	157.4	4.4	17.3	1.8	6.7	301
	May	205.8	0.0	21.5	2.1	7.6	335
	June	216.2	0.0	26.3	2.0	12.1	337
	July	224.1	0.0	29.1	1.8	14.1	331
	August	197.4	0.0	27.2	1.6	14.1	303
	September	146.6	1.6	23.9	1.5	12.7	248
	October	88.8	35.4	15.9	1.4	8.5	178
	November	40.1	6.9	11.8	1.1	9.9	112
	December	22.1	92.9	8.1	1.5	6.9	73
2022	January	40.8	2.2	7.8	1.0	6.2	120
	February	67.5	3.8	9.6	1.3	2.2	182
	March	103.4	34.9	13.7	1.6	6.3	215
	April	146.8	5.4	16.3	2.0	6.3	286
	May	201.3	0.0	20.1	2.3	7.1	334
	June	217.9	1.3	25.4	2.2	10.3	342
	July	222.8	0.0	27.7	1.9	13.0	327
	August	197.8	2.5	27.8	1.7	13.9	295
	September	147.7	3.1	25.1	1.6	13.7	242
	October	102.7	0.0	18.3	1.3	10.3	194
	November	49.8	4.1	8.2	1.3	4.0	133
	December	19.8	74.2	7.2	1.5	5.7	67

Source: Data from the State of California (n.d.).

Soil Matric Potential Measurements

Watermark soil moisture sensors (Irrometer, Riverside, California Irrometer, n.d.) were used to monitor soil water status by measuring soil matric potential (in kPa) throughout the soil profile at different depths (locations are shown in Fig. 1). The moisture sensors were installed in all strips for each of the two irrigation treatments (Full/Deficit) at four depths: 30, 60, 90, and 120 cm to cover the active root zone (depth of maximum root intensity) and account for soil water distribution. Soil moisture data were recorded hourly using a Watermark 900M monitor with 8 sensor capacity. Data were downloaded using a 900DS data shuttle or laptop and then converted to the daily averaged matric potential.

Forage Yield, Irrigation Water Productivity, and Crop Water Productivity

The impact of deficit irrigation on alfalfa field quality was considered in this study. Three alfalfa samples were collected from each plot during cuttings for the entire duration of the study. In the Central Valley of California, alfalfa is typically cut 7 or 8 times per season. Alfalfa was harvested every 28–30 days with a total of seven cuts in the first year (2019) and eight cuts for each of 2020, 2021, and 2022. A plot harvesting machine (Carter, Brookstone, Indiana) was used to cut the alfalfa perpendicular to the strips’ length in three locations per strip, with 0.9 m wide by 5.4 m long samples collected for yield assessment. Yield per unit area was calculated based on total fresh weight multiplied by dry matter concentration.

Two Tule Technologies systems (Tule Technologies Inc., Davis, California) (which is a proprietary commercialized version of surface renewal methods, currently only available in California) were used to determine the alfalfa actual evapotranspiration (

E T_{a}

) (Tule Sensors 2024). The Tule system is a commercial version of the surface renewal method based on the Ph.D. thesis of Dr. Tom Shapland who conducted his work at UC Davis and resulted in a patented approach to estimate

E T_{a}

based on the surface renewal method without the need to calibrate with Eddy covariance. Daniele Zaccaria of UC Davis compared this method to the standard Eddy covariance measurement on alfalfa and the Tule Technologies system results were within 95% confidence level. As far as the fetch, since the Tule system is relatively short (about 0.6–0.75 m above the ground), Tule Technologies has a quality assurance protocol that determined that the plot dimensions in this study were within the confidence level that they have.

The Tule Technology estimates

E T_{a}

based on the principles of the surface renewal (SR) method (Shapland et al. 2014). The SR is a biometeorological method that uses high-frequency air temperature measurements above the plant canopy to estimate sensible heat flux. The sensible heat flux is then used, along with net radiation and soil heat fluxes, to estimate latent heat flux as the residual in the surface energy balance equation. The Tule Technologies patented system eliminates the need for calibration of SR against other methods such as Eddy covariance to obtain estimates of sensible heat flux, and this need for calibration limited the use of SR to research applications (Shapland et al. 2014). Therefore, irrigation decisions during this study were based on the crop water requirements as estimated by Tule Technologies stations. Additional information about this relatively new commercial application of the SR method for measuring actual evapotranspiration can be found on the Tule Technologies website (Tule Sensors 2024).

Irrigation water productivity (IWP in

Kg {ha}^{- 1} {mm}^{- 1}

) was calculated for each treatment as the ratio of alfalfa hay yield (

kg {ha}^{- 1}

) to applied irrigation water (mm) (Howell et al. 1990) [Eq. (1)]. IWP is used in this study as a generalized measure of alfalfa biomass produced per unit amount of water applied. The parameter encompasses the irrigation and scheduling method and soil conditions, among many factors that could affect plant growth. However, the imposed differences in irrigation amounts were the only known controlled variables that could be directly related to IWP. Crop water productivity (CWP) for each harvest period was calculated as the ratio of yield (

kg {ha}^{- 1}

) to the

E T_{a}

(mm) [Eq. (2)]

I W P = \frac{Y i e l d}{I r r i g a t i o n a p p l i e d w a t e r}

(1)

C W P = \frac{Y i e l d}{A c t u a l e v a p o t r a n s p i r a t i o n}

(2)

where yield is in

kg {ha}^{- 1}

, and irrigation AW and actual evapotranspiration,

E T_{a}

, is in mm.

Statistical Analysis

Statistical analyses were performed using IBM SPSS statistical software (IBM, n.d.) to analyze and compare the difference in alfalfa yield between the full irrigation and deficit irrigation treatments. The statistical analysis was performed as follows: (1) a single sample T-test was executed for yield among cuts for each treatment, and each year separately (two-tailed hypothesis,

p < 0.05

) to explore the difference in yield among cuts, (2) one-way ANOVA test was performed for the two independent treatments (full, and deficit) for each year to compare the means of yield between the full and deficit irrigation treatments for each year of the experiment, and (3) run ANOVA test between years (as repeated measures) for the full and deficit treatments to identify if the yield (for full and deficit treatments) was significantly reduced throughout years. Additionally, a regression analysis was performed to develop relationships between yield (for each year) and AW (for full and deficit treatments) and between the cumulative yield of each cutting and cumulative evapotranspiration, where the coefficient of determination,

R^{2}

, was utilized to judge the goodness of fit of these relationships.

Results and Discussions

Actual Evapotranspiration

The actual evapotranspiration (

E T_{a}

) of each cutting is presented in (Table 3). The

E T_{a}

and

K_{c}

exhibited a similar trend for both fully and deficit-irrigated treatments (Fig. 3).

E T_{a}

increased to its maximum values (

7.0 - 7.5 mm {day}^{- 1}

) by the end of June followed by a decrease toward the end of the season.

K_{c}

values were obtained by dividing the

E T_{a}

by

E T_{o}

. The average observed

K_{c}

values for the full irrigation treatment, excluding values before the first cut and after the last harvest, were 0.77, 0.73, 0.75, and 0.76 for 2019, 2020, 2021, and 2022, respectively. The

K_{c}

values for the midsummer deficit irrigation treatment were 0.77, 0.72, 0.74, and 0.72, for 2019, 2020, 2021, and 2022, respectively, which was smaller than the reported value of 0.99 by Hanson et al. (2007) for the Sacramento Valley. A small difference was found in

E T_{a}

values between the full and deficit irrigation treatments. This may be explained by the deep root system of alfalfa and the availability of water in the soil profile (Bai and Li 2003).

Table 3. Cutting cycles for each season showing the AW and

E T_{a}

for the full and deficit treatments, rainfall, and percentage of water saving

No.	Cycle		AW (mm)		Rainfall (mm)	$E T_{a}$ (mm)		Water savings (mm)	Water savings (%)
No.	Start	End	Full	Deficit	Rainfall (mm)	Full	Deficit	Water savings (mm)	Water savings (%)
2019
1st	1st January	6th May	0	0	147	298	291	454	30.3
2nd	6th May	3rd June	175	232	54	125	127
3rd	3rd June	2nd July	278	279	0	173	174
4th	2nd July	1st August	275	267	0	180	180
5th	1st August	29th August	268	264	0	185	180
6th	29th August	2nd October	247	0	0	136	144
7th	2nd October	14th November	253	0	0	75	79
Summation			1,496	1,042	201	1,172	1,175
2020
1st	1st January	24th April	0	0	114	259	267	706	40.9
2nd	24th April	26th May	113	147	8	163	145
3rd	26th May	24th June	333	284	0	172	168
4th	24th June	22nd July	312	297	0	160	158
5th	22nd July	18th August	288	291	0	140	136
6th	18th August	14th September	248	0	0	110	104
7th	14th September	13th October	231	0	0	80	76
8th	13th October	16th November	200	0	5	44	49
Summation			1,725	1,019	127	1,128	1,103
2021
1st	1st January	7th April	0	0	88	213	210	625	37.0
2nd	7th April	5th May	238	317	5	129	122
3rd	5th May	8th June	224	273	0	199	198
4th	8th June	7th July	238	237	0	178	176
5th	7th July	4th August	248	237	0	163	160
6th	4th August	1st September	246	0	0	136	128
7th	1st September	29th September	275	0	2	107	101
8th	29th September	28th October	220	0	43	59	53
Summation			1,689	1,064	138	1,184	1,148
2022
1st	1st January	7th April	0	0	41	191	182	815	49.1
2nd	7th April	25th May	293	297	5	231	226
3rd	25th May	22nd June	301	295	1	156	150
4th	22nd June	20th July	269	253	0	155	155
5th	20th July	17th August	239	0	3	136	129
6th	17th August	15th September	244	0	0	116	103
7th	15th September	12th October	220	0	3	76	64
8th	12th October	10th November	94	0	0	39	35
Summation			1,660	845	53	1,100	1,044

Fig. 3. Actual evapotranspiration ( $E T_{a}$ ), reference evapotranspiration ( $E T_{o}$ ), and daily crop coefficient ( $E_{c}$ ) on a weekly basis for (a) fully irrigated treatment; and (b) Midsummer deficit irrigation.

The cumulative

E T_{a}

for the fully irrigated treatment was 1,172, 1,128, 1,184, and 1,100 mm; and the cumulative

E T_{a}

for the deficit-irrigated treatment was 1,175, 1,103, 1,148, and 1,044 mm in 2019, 2020, 2021, and 2022, respectively (Fig. 4). This small difference in

E T_{a}

between fully and deficit-irrigated treatments demonstrates that

E T_{a}

cannot be a reliable approach to calculate the amount of water saved to transfer to the low water availability areas as it is a site-specific response. The cumulative

E T_{a}

for the fully irrigated treatments was lower than the AW, meaning that the fully irrigated plots were overirrigated throughout the study period where surface irrigation of deep-rooted crops can be very efficient in the SJV. Seasonal AW exceeded

E T_{a}

by 324, 597, 505, and 560 mm for 2019, 2020, 2021, and 2022, respectively, for the fully irrigated treatment. Seasonal AW for the midsummer deficit treatment was lower than the crop water requirements where the difference between

E T_{a}

and the AW during deficit were 133, 84, 84, and 199 mm. Cumulative

E T_{a}

of alfalfa is almost the same for full and deficit treatments as alfalfa roots developed over time and a substantial percentage of total

E T_{a}

can come from deeper soil where moisture is relatively high or from groundwater contribution (3 m-depth) Lindenmayer et al. (2011).

Fig. 4. Cumulative actual evapotranspiration for fully and deficit-irrigated treatments and the cumulative average reference evapotranspiration over the study period.

Soil Water Potential

Soil matric potential values (in kPa) for full and deficit irrigation treatments were averaged daily and shown in Figs. 5(a and b). The data clearly showed the irrigation events for the full/deficit treatments throughout the entire experiment, where soil matric potential was suddenly increased after water applications and gradually decreased until the sequent irrigation event. The minimum soil matric potential values were kept in the range between

- 40

and

- 50 kPa

(for both treatments) for alfalfa growing in sandy loam soil, as recommended by Orloff et al. (2005). Soil matric potential values were approximately

- 50 kPa

before harvest and increased to approximately

- 5 kPa

after applying irrigation.

Fig. 5. Daily average soil matric potential (kPa) at different depths of 30 cm, 60 cm, 90 cm, and 120 cm under (a) fully irrigated treatment; and (b) midsummer deficit-irrigated treatment.

The soil matric potential readings remained relatively high between the consecutive irrigation events in each harvest period. Soil matric potential was high (high water content) in the deeper soil profile, while soil matric potential in the top 30 cm layer fluctuated between dry and wet, which was greatly affected by irrigation treatments. The shallow layer potential is relatively low (shown in the blue line) where the soil is dry for the deficit treatment [Fig. 5(b)] as compared to the full irrigation treatment [Fig. 5(a)]. This is consistent with the other findings from irrigated alfalfa studies (Mundy et al. 2006; Huang et al. 2018; Wang et al. 2021). Soil matric potential reached its highest value (

- 200 kPa

) gradually after imposing deficit irrigation in late August, indicating that little water was available through the soil profile to the alfalfa which is similar to the finding of Pembleton et al. (2011). The greatest observed changes in soil water potential were in the layers 0–30 cm and 30–60 cm as those layers dried faster than the deeper layers. However, there were some responses to soil matric potential in the other layers as well.

Applied Irrigation Water

For the year preceding this research experiment (2018), an equal amount of irrigation was applied across all the plots. Two irrigation events were applied between each cutting during the following four years of the study (2019 to 2022). The deficit irrigation treatments were implemented on 14th August, 31st August, 17th August, and 1st August of years 2019, 2020, 2021, and 2022, respectively. The total amount of applied IW for the fully irrigated treatment was 1,496, 1,725, 1,689, and 1,660 mm in 2019, 2020, 2021, and 2022, respectively, while it was 1,042, 1,019, 1,064, and 845 mm during 2019, 2020, 2021, and 2022, respectively, for the deficit-irrigated treatment (Fig. 6).

Fig. 6. Cumulative applied IW to the fully and midsummer deficit-irrigated treatments throughout the four years.

IW savings due to deficit treatments were 30.3%, 40.9%, 37.0%, and 49.1%; while the reductions in the AW between full and deficit treatments were 454, 706, 625, and 815 mm in 2019, 2020, and 2021, respectively (Table 3), which is of a great benefit during drought years. This highlights the importance of midsummer deficit irrigation in water conservation that can be made available to transfer to areas of water shortage, a water conservation practice that was recommended by Orloff et al. (2014). The amount of rainfall during the seasons of 2019, 2020, 2021, and 2022 for alfalfa in the SJV were 201, 127, 138, and 53 mm, respectively. Leaching was not considered during these study periods and the average salinity of IW was

0.24 dS / m

which was well below the

2.0 dS / m

soil salinity threshold (ECe) of alfalfa (Leinfelder-Miles 2013). The salinity of the soil profile was below

2.0 dS / m

during the entire experiment. The groundwater that is used as the source for irrigation in the region is recharged by the Kings River water which is only 3 Km to the east of the experimental site and fed by runoff from snowmelt from the Sierra Nevada mountains with relatively low salinity. The overall irrigation efficiency during this study period is comparable to the application efficiency of alfalfa grown in the region (Hanson and Putnam 2004). The overall efficiency in 2019 was 78% (

E T_{a}

/AW or

1,172 / 1,496

), 65% in 2020, 70% in 2021, and 66% in 2022 (Table 3). These efficiency numbers are comparable to application efficiencies in commercial fields with relatively light soil (Hanford sandy loam).

Hay Quality and Dry Matter Yield

There were no significant differences in most hay quality parameters within any of the years between 2019 and 2022. The deficit irrigation treatments resulted in a small but not significant decline or improvement in hay quality parameters and had no impact on hay quality classification. The overall impact of deficit irrigation on most individual hay quality parameters was negligible and, in some cases, improved one or more of the quality parameters [for example Neutral detergent fibers (NDF) in 2019, 2020, and 2021].

The quality of the hay for all irrigation treatments for all years was good and classified as premium, premium-good, or supreme with excellent retail price (Table 4). The premium hay classification is given to alfalfa that is cut prebud, bud, or early bloom with low fiber, soft stems, high energy and intake potential, and good leaf attachments (Putnam and Orloff 2016). Premium alfalfa is also mostly free of grasses and weeds, no noxious weeds, no mold, and well cured.

Table 4. Effect of deficit irrigation treatments on forage nutritive value parameters and market classification due to quality (USDA-Agricultural Marketing Service designations)

Year	Treatment	NDF	CP	ADF	Lignin	Hay quality classification
${g kg}^{- 1}$
2019	Full	348	245	289	55	Premium
2019	Deficit	341	236	278	54	Premium
2020	Full	332	256	281	54	Premium
2020	Deficit	329	257	276	54	Premium
2021	Full	380	261	297	73	Premium-Good
2021	Deficit	377	259	294	72	Premium-Good
2022	Full	311	260	262	57	Supreme
2022	Deficit	311	251	258	58	Supreme

Note: NDF (neutral detergent fibers), CP (crude Protein), and ADF (acid detergent fibers).

Yield responses to irrigation treatments of each harvest cycle for the four years of the study are presented in Table 5 and the contribution as a percentage of each cut’s yield to the total yield is presented in Figs. 7(a and b) for the full and deficit treatments, respectively. Previous studies found that the dry matter (DM) yield of alfalfa was affected by irrigation treatment (Qiu et al. 2021) and how much water was applied relative to

E T_{a}

(Putnam et al. 2017b). Analysis of these yield results showed that there was no significant difference between yields for full and deficit treatments for the first five cuttings, while the deficit treatment had significantly lower yields for later cuttings among all four years [Figs. 8(a–d)]. Also, a significant difference was obtained between the yearly cumulative yield for the two treatments (T-test, at a significance

level = 0.05

) (Fig. 9).

Table 5. Alfalfa yield (

Mg {ha}^{- 1}

) from individual cuttings, total yield, and yield reduction for the deficit treatment relative to the full irrigation treatment

Cuts	1st	2nd	3rd	4th	5th	6th	7th	8th	Total $\pm$ STD ( $σ$ )	Yield reduction ( $Mg {ha}^{- 1}$ )
2019
Harvest date	6th May	3rd June	2nd July	1st August	29th August	2nd October	14th November	—	—	—
Full	3.22	4.14	3.50	3.43	3.11	2.05	2.29	—	$21.75 \pm 0.72$	3.94
Deficit	3.25	3.72	3.27	3.20	3.11	0.80	0.45	—	$17.81 \pm 1.33$	3.94
2020
Harvest date	24th April	26th May	24th June	22nd August	18th August	14th September	13th October	16th November	—	—
Full	3.62	2.77	3.34	3.22	1.96	1.23	1.31	1.01	$18.47 \pm 0.99$	2.04
Deficit	3.58	2.42	3.18	3.35	1.88	0.79	0.94	0.28	$16.43 \pm 1.14$	2.04
2021
Harvest date	7th April	5th May	8th June	7th July	4th August	1st September	29th September	28th October	—	—
Full	3.24	2.41	3.03	3.11	2.11	1.50	1.55	0.68	$17.63 \pm 0.73$	1.25
Deficit	3.17	2.35	3.35	3.34	2.18	1.13	0.64	0.24	$16.38 \pm 1.08$	1.25
2022
Harvest date	7th April	25th May	22nd June	20th July	17th August	15th September	12th October	10th November	—	—
Full	2.37	1.84	2.22	2.79	1.53	1.41	1.00	0.34	$13.51 \pm 0.62$	0.40
Deficit	2.85	2.28	2.70	3.31	1.25	0.42	0.12	0.18	$13.11 \pm 1.25$	0.40

Fig. 7. Contribution of each cut to the total yield among the four years of study for (a) full irrigation; and (b) deficit irrigation.

Fig. 8. Yield at each cut under full and deficit-irrigated treatments for (a) 2019; (b) 2020; (c) 2021; and (d) 2022.

Fig. 9. The total yearly and average yield for the fully and deficit-irrigated treatments and vertical bars are standard errors of the mean values.

Table 6 summarizes the statistical results for the single T-test among cuttings of each treatment for each year. There was no significant difference in yield among cuts, either for fully irrigated treatments or deficit irrigation for the four years. Statistical analysis also revealed that there is no significant difference between the yield of the fully irrigated treatments and the yield of the deficit-irrigated treatments for any year (

ρ

-values are summarized in Table 6). However, a significant difference in yield over the years (as four repeated measures) was obtained for the fully irrigated and deficit-irrigated treatments where

ρ - value = 0.0001

, 0.0129, respectively.

Table 6. Descriptive analysis and single T-test results between cuts, ANOVA test between full and deficit treatments for each year, and ANOVA test between years for full and deficit treatments

Year	Single T-test					ANOVA test (between full and deficit)		ANOVA test (between years) for full and deficit
Year		Mean	Median	$t$ -value	$ρ$ -value	$F$ -ratio	$ρ$ -value	(Full irrigation)	(Deficit irrigation)
2019 (7 cuts)	Full	3.11	3.22	$- 0.0157$	0.4939	0.9706	0.3439	$F - ratio = 12.8724$	$F - ratio = 4.5747$
2019 (7 cuts)	Deficit	2.54	3.20	0.0057	0.4978	0.9706	0.3439	$F - ratio = 12.8724$	$F - ratio = 4.5747$
2020 (8 cuts)	Full	2.31	2.37	0.0201	0.4923	0.1894	0.6701	$ρ - value = 0.0001$ *	$ρ - value = 0.0129$ *
2020 (8 cuts)	Deficit	2.05	2.15	0.0055	0.4979	0.1894	0.6701
2021 (8 cuts)	Full	2.20	2.26	0.0116	0.4956	0.0793	0.7824
2021 (8 cuts)	Deficit	2.05	2.27	$- 0.1137$	0.4564	0.0793	0.7824
2022 (8 cuts)	Full	1.69	1.69	$- 0.0089$	0.4966	0.0082	0.9292
2022 (8 cuts)	Deficit	1.64	1.77	$- 0.0027$	0.4989	0.0082	0.9292

Note: (*) means that the difference is significant at $p < 0.05$ .

The average yearly alfalfa DM yields differed over the years. It was highest in the first year of production (2019), achieving 21.75 and

17.81 Mg {ha}^{- 1}

for fully and deficit-irrigated treatments, respectively. These values of the DM yields coincide with the study of Mengistu et al. (2022) who reported that the mean value of DM for four alfalfa cultivars during two cuts in 2016 and three cuts in 2017 at Masha Highland was 6.3 and 5.9 (

Mg {ha}^{- 1}

12.2 in-total), respectively. Feng et al. (2022) also found similar results where the average DM yield of alfalfa was

11.18 \pm 6.69 Mg {ha}^{- 1}

. Alfalfa during 2019 was cut seven times with an 18% reduction in yield for deficit irrigation compared to the fully irrigated treatment. 2019 had a greater reduction in yield for the deficit-irrigated treatment, than the other sequent three years.

Yields were 21.75, 18.47, 17.63, and

13.51 Mg {ha}^{- 1}

for fully irrigated treatments and 17.81, 16.43, 16.38, and

13.11 Mg {ha}^{- 1}

for the deficit irrigation for 2019, 2020, 2021, and 2022, respectively. Alfalfa can be maintained for four to five years, sometimes longer (for disease-resistant varieties) depending on cutting management. The yield still tends to decline after the second production year due to damage from wheel traffic (causing topsoil compaction), winter diseases in older stands, and long-term soil moisture depletion (Beckman 2022). Thus, results clearly indicate that planning shorter rotations or replanting alfalfa after the third year in production rather than keeping a low-yielding field is strongly recommended, in terms of profitability, which is similar to the observations of Capstaff and Miller (2018). These results concur with Undersander who reported that the alfalfa trials have been highest in the first and second production years and then began to decline with 15% as an average in the third production year and almost 30% in the fourth production year (Undersander 2001).

Irrigation Water Productivity and Crop Water Productivity

The relationship between the applied IW and yield was analyzed for four years for the full and deficit irrigation treatments (Fig. 10). The generated relationship between AW and yield is the IW production function based on four-year experimental research of alfalfa. 2nd-order polynomial functions were correlated with the relationship between alfalfa yield and applied IW for both full and deficit irrigation treatments. A similar trend was observed by Djaman et al. (2020) where alfalfa forage yield was significantly affected by the irrigation regimes and showed a third-order polynomial relationship with the applied irrigation amount. The coefficient of correlation,

R^{2}

is 0.89 for the two equations for both treatments.

Fig. 10. Yield ( $kg {ha}^{- 1}$ ) as a function of applied IW (mm) as an average of the four years.

IWP and CWP values are summarized in Table 7. The deficit-irrigated treatments have a considerably higher IWP as compared to the fully irrigated treatments. IWP was 14.54, 10.71, 10.44, and

8.14 kg {ha}^{- 1} {mm}^{- 1}

for the fully irrigated treatment and was 17.09, 16.11, 15.40, and

15.54 kg {ha}^{- 1} {mm}^{- 1}

for the deficit treatment for 2019, 2020, 2021, and 2022, respectively. This is because the applied IW for deficit treatment is much lower than the irrigation AW for the full irrigation treatment. Our finding shows that alfalfa has higher IWP and is more tolerant to deficit irrigation than many perennial forage crops (Kelly et al. 2006; Neal et al. 2011). Since the first cutting cycle (of each year) did not receive water from irrigation where the crop was dependent only on rainfall, IWP for this cycle is the highest among all cutting cycles; however, results for IWP as well as CWP are presented as average for the whole year. CWP was compared between the full irrigation and deficit treatment for the four years of study where it resulted in 18.59, 16.37, 14.91,

12.28 kg {ha}^{- 1} {mm}^{- 1}

for full irrigation and 15.17, 14.91, 14.27, and

12.56 kg {ha}^{- 1} {mm}^{- 1}

for deficit treatment for 2019, 2020, 2021, and 2022, respectively. The summary of the database (68 manuscripts) for alfalfa water productivity and DM yield is summarized in Table 1 of Fink et al. (2022).

Table 7. Yield, IWP, and crop water productivity (CWP) for full and deficit treatments

No.	Cycle		Yield ( $kg {ha}^{- 1}$ )		IWP ( $Kg {ha}^{- 1} {mm}^{- 1}$ )		CWP ( $Kg {ha}^{- 1} {mm}^{- 1}$ )
No.	Start	End	Full	Deficit	Full	Deficit	Full	Deficit
2019
1st	1st January	6th May	3,220	3,250	14.54	17.09	18.59	15.17
2nd	6th May	3rd June	4,143	3,722
3rd	3rd June	2nd July	3,508	3,272
4th	2nd July	1st August	3,430	3,197
5th	1st August	29th August	3,110	3,110
6th	29th August	2nd October	2,049	801
7th	2nd October	14th November	2,292	454
Summation			21,752	17,806
2020
1st	1st January	24th April	3,622	3,584	10.71	16.11	16.37	14.91
2nd	24th April	26th May	2,771	2,424
3rd	26th May	24th June	3,342	3,179
4th	24th June	22nd July	3,223	3,355
5th	22nd July	18th August	1,960	1,876
6th	18th August	14th September	1,229	794
7th	14th September	13th October	1,308	938
8th	13th October	16th November	1,012	276
Summation			18,466	16,427
2021
1st	1st January	7th April	3,238	3,165	10.44	15.40	14.91	14.27
2nd	7th April	5th May	2,407	2,351
3rd	5th May	8th June	3,031	3,345
4th	8th June	7th July	3,113	3,337
5th	7th July	4th August	2,106	2,183
6th	4th August	1st September	1,503	1,126
7th	1st September	29th September	1,549	641
8th	29th September	28th October	683	235
Summation			17,630	16,384
2022
1st	1st January	7th April	2,372	2,852	8.14	15.54	12.28	12.56
2nd	7th April	25th May	1,844	2,283
3rd	25th May	22nd June	2,219	2,704
4th	22nd June	20th July	2,790	3,307
5th	20th July	17th August	1,533	1,247
6th	17th August	15th September	1,413	422
7th	15th September	12th October	1,001	121
8th	12th October	10th November	341	177
Summation			13,513	13,113

The relationship between the cumulative evapotranspiration (mm) and cumulative yield (

kg {ha}^{- 1}

) for the two irrigation treatments (full irrigation, and the midsummer deficit irrigation) throughout the four years is shown in Figs. 11(a–d). Cumulative evapotranspiration is similar for the four years and within irrigation treatments (full and deficit). A linear relationship between cumulative yield and cumulative evapotranspiration (for the midsummer deficit irrigation treatment) was found with a good correlation where

R^{2}

was 0.96, 0.98, 0.98, and 0.99 for 2019, 2020, 2021, and 2022, respectively.

Fig. 11. Correlation between cumulative yield ( $kg {ha}^{- 1}$ ) and cumulative actual evapotranspiration (mm) for (a) 2019; (b) 2020; (c) 2021; and (d) 2022.

Conclusions

Generally, when water resources are limited and drought conditions limit water availability for irrigation, efficient irrigation management practices are essential for achieving water conservation. Alfalfa is the major field crop in California, and midsummer deficit irrigation could be used to generate a significant amount of water to alleviate the impact of drought. This study documented the impact of midsummer deficit irrigation (for four years) on IWP and crop water productivity (CWP) as compared to normal irrigation practices (full irrigation).

Alfalfa yield was higher in the first year and significantly declined over time. There was no significant difference in yield between full and deficit treatment for the four years. The net IW savings were 454, 706, 625, and 815 mm in 2019, 2020, 2021, and 2022 that represented 30.3%, 40.9%, 37.0%, and 49.1% of the AW. In the San Joaquin Valley of California, midsummer deficit-irrigated treatments had a higher IWP than those fully irrigated treatments.

CWP for the full and deficit treatment was almost the same for each year since crop evapotranspiration and yield were almost the same as well. These results suggest that IWP may be more “precise,” but it is not more useful, as water conservation depends upon consumed water (

E T_{a}

). Seasonal IWP is a more precise tool than seasonal CWP to study the plant-water relationship. Additionally, crop production function was identified using applied IW. Our finding revealed that midsummer deficit irrigation practices could be implemented and considered by alfalfa growers during dry years when the availability of IW is limited. Further work is needed to evaluate the cost-benefit and the tradeoff between yield loss and water savings including energy and labor costs.

Data Availability Statement

Some or all data, models, or code generated or used during the study are available from the corresponding author by request (CIMIS, TULE, and yield data).

Acknowledgments

This work was supported by UC Kearney Agricultural Research and Extension Center and USDA-Agricultural Research Service (ARS) Collaborative Agreement No. 58-2034-8-038. This research was also supported by USDA Natural Resources Conservation Agreement NR183A750023C005.

References

Alam, M., T. P. Trooien, T. J. Dumler, and D. H. Rogers. 2002. “Using subsurface drip irrigation for alfalfa 1.” JAWRA J. Am. Water Resour. Assoc. 38 (6): 1715–1721. https://doi.org/10.1111/j.1752-1688.2002.tb04376.x.

Abstract

Introduction

Material and Method

Study Area and Experiment Design

Climate Data

Soil Matric Potential Measurements

Forage Yield, Irrigation Water Productivity, and Crop Water Productivity

Statistical Analysis

Results and Discussions

Actual Evapotranspiration

Soil Water Potential

Applied Irrigation Water

Hay Quality and Dry Matter Yield

Irrigation Water Productivity and Crop Water Productivity

Conclusions

Data Availability Statement

Acknowledgments

References

Information

Published In

Copyright

History

ASCE Technical Topics:

Authors

Affiliations

Metrics

Citations

Download citation

Figures

Other

Share

Copy the content Link

Share with email

Share

Request Username

Create a new account

Change Password

Password Changed Successfully

Verify Phone

Congrats!